MAGENTA (Making Genetic testing accessible): a prospective randomized controlled trial comparing online genetic education and telephone genetic counseling for hereditary cancer genetic testing

Online access to physician ordered genetic testing for clinically actionable hereditary cancer genes has the potential to increase access to genetic testing while maintaining test quality and clinical utility. Smit and colleagues found that individuals believe online communication is an acceptable mode of receiving medical information, and that it is important to have a health professional available to discuss questions and content as needed [6]. One study has examined the use of online genetic testing for cascade testing of family members and found it to be an effective approach to implementing cascade testing, citing an uptake of cascade testing in 47.5% of first degree relatives [7].

The availability of online genetic education detailing the benefits and limitations of hereditary cancer testing with access to board certified genetic counselors offers a promising alternative that may overcome barriers faced by current SDMs. Such a model could allow both genetics and non-genetics clinicians to order genetic testing while providing patients with detailed online information regarding hereditary cancer testing, overcoming the major barriers of physician coordinated testing, including limitations of physician time and knowledge. Remote access testing may eliminate the major barriers of in-person counseling, including time and money spent on travel, while increasing flexibility and accessibility, which continue to be challenges of both in-person counseling and telephone counseling. Remote access genetic testing via the use of an online genetic services platform allows individuals to access information using the internet with the ability to revisit if they choose, avoiding the need for appointments and allowing for patient-driven and convenient genetic services. Electronic copies of genetic test results may be shared easily with family members and medical providers. For individuals found to have a pathogenic mutation, online platforms facilitate easy information sharing, and the option of online access to testing services may promote the uptake of cascade testing in blood relatives.

One question in this proposed SDM is how much personalized counseling is needed both pre-test and post-test to optimize patient outcomes. Does required personalized counseling reduce patient distress and improve understanding or does it add unnecessary cost and time and reduce likelihood of test completion? Is pre- or post-test counseling more important and if counseling is provided as an option, how many patients will utilize telephone counseling, and can we identify patient subsets who have greater benefit from personalized counseling? To address these considerations, we are conducting a prospective study to evaluate different amount of personalized genetic counseling within an online genetic testing model to determine less counseling is non-inferior to our current accepted SDM at delivering hereditary cancer risk information. The purpose of this paper is to describe the design of the MAGENTA study, a prospective randomized trial of online genetic education vs telephone genetic counseling for hereditary cancer genetic testing.

The MAGENTA (MAking GENetic Testing Accessible) trial is designed to compare the effectiveness of pre- and post-test telephone counseling to three potentially more accessible alternatives for individuals undergoing analysis of 19 hereditary breast and ovarian cancer (OC) genes. The genes included are: ATM, BARD1, BRCA1, BRCA2, BRIP1, CDH1, CHEK2, EPCAM, MSH2, MSH6, MLH1, NBN, PALB2, PMS2, PTEN, RAD51C, RAD51D, TP53, and STK11. Our primary hypothesis is that the outcomes for subjects receiving one of the three alternative approaches utilizing genetic education online will not be inferior to those of participants receiving a current acceptable standard of telephone genetic counseling both pre- and post-testing [4, 5].

After online eligibility screening and electronic study consent (as approved by the Human Subjects Division of the MD Anderson Institutional Review Board), randomization into one of four study arms is performed automatically using a web-based data management platform [8]. The four study arms were selected to independently assess the delivery of genetic information both before and after genetic testing (pre-test and post-test) by either telephone genetic counseling or online education alone. The control arm requires mandatory pre and post-test telephone counseling while the other three arms provide various combinations (Table 1). Participants are mailed a saliva collection kit after all pre-test online education and/or genetic counseling is completed, based on study arm. Analysis of 19 OC genes is completed once the participant returns the saliva sample. As an additional safety measure, any participant found to carry a pathogenic or likely pathogenic mutation has required post-test telephone counseling and any participant desiring genetic counseling is provided with telephone genetic counseling regardless of study arm. A study schema can be found in Fig. 1. Participants were stratified into one of two risk groups: individuals with a personal and/or family history of cancer, and those with a known familial mutation (cascade group).

The MAGENTA study’s target population is women that may be at increased risk to develop ovarian cancer based on their personal or family history of cancer or due to a known cancer susceptibility gene mutation in the family. Women must be 30 years of age or older and must have at least one ovary, as we hope to identify women who will be able to take actionable measures to reduce ovarian cancer risk. As a safety precaution, all participants are required to identify a local healthcare provider with whom their genetic test results will be shared. Complete eligibility criteria are outlined in Table 2. Approximately 5200 women will be enrolled with a goal of 3000 completing genetic testing. We estimate that we will need to screen between 7000 and 10,000 women using an online eligibility questionnaire.

Recruiting efforts are being carried out through clinical settings, advocacy groups, and various social media platforms. The different recruitment strategies were designed to capture a heterogeneous group of women, which will allow us to assess whether we improve access to genetic testing in diverse patient populations stratified for race and ethnicity, as well as, socioeconomic, educational and geographical backgrounds.

Given the size and large geographic area (all 50 U.S. states and the District of Colombia) for which we are recruiting and enrolling participants, the use of electronic informed consent (IC) became a vital component to the study. Electronic informed consent is defined by the FDA as the use of electronic systems and processes that may employ multiple electronic media, such as text, graphics and videos, to convey information related to the study and obtain and document informed consent (http://​www.​fda.​gov/​downloads/​drugs/​guidancecomplian​ceregulatoryinfo​rmation/​guidances/​ucm436811.​pdf).

Our biggest challenge was ensuring that the person signing the IC is the same as the person completing the study procedures. We accomplished this by requiring all participants to create a unique username and password that must be used to confirm their identity before signing the IC and before completing each study survey. To ensure that consent is truly informed, we created a series of comprehension questions that the participant must answer correctly before being allowed to sign the IC. The IC was delivered to study participants by email using the survey tools in REDCap. All other aspects of the FDA draft guidance were followed, including allowing the participant the ability to print a copy of the IC and call the study coordinator for questions (http://​www.​fda.​gov/​downloads/​drugs/​guidancecomplian​ceregulatoryinfo​rmation/​guidances/​ucm436811.​pdf).

To fully evaluate the efficacy of these alternate models, we are assessing several different outcome measures, including: cancer-risk distress, testing completion rate (defined as receiving results), genetics knowledge, cognitive outcomes (i.e. decisional satisfaction, decisional regret), and behavioral outcomes. Surveys are completed at baseline, 3 months, 12 months and 24 months post-results disclosure. The specific measures and time-points are summarized in Table 3. These measures were informed by the prior work of Kinney et al. and Schwartz et al. [4, 5]. Following the Peshkin et al. method of classifying these measures, these variables were chosen to assess cognitive (decisional regret and decisional satisfaction), affective (distress, anxiety, depression), and behavioral outcomes [18]. Questionnaires for assessing outcome measures are delivered to study participants by email using the survey features of REDCap. Questionnaires are distributed in an automated fashion, based on the date of the participant’s results disclosure.

The primary study outcome is cancer-risk distress at 3 months post-result disclosure, as measured by the Impact of Events Scale (IES) [9, 19].

All participants receive online genetic education, including a 5-min pre-test educational video and an online results report. The video provides an explanation of hereditary cancer, inheritance patterns, associated cancer risks, and an overview of the possible results. Participants randomized to receive pre-test telephone counseling are prompted to schedule their telephone appointments online and will complete telephone counseling in addition to the online genetic education. The usual content of pre- and post-test genetic counseling for hereditary cancer risk has been well described [20, 21]. Following these standard guidelines, a genetic counseling outline was created by the study team and is followed by the board-certified genetic counselors involved in the study to guarantee consistency in counseling and ensure standardization for the purposes of the study (Table 4).

We employ a non-inferiority design with a non-inferiority margin of 4 points, as suggested by Schwarz et al. (2014), for the difference in IES score between the control arm and each of the other three study arms at 3 months post-result disclosure.

We are enrolling 5200 eligible women with a goal of 3000 completing genetic testing. We expect that around 85% (n = 4400) of those who consent to participate will complete the baseline questionnaire and then be randomized. Of these, we will enroll around 750 subjects with a known familial mutation into the cascade group, with the remaining 3650 subjects, who have an increased likelihood of carrying a deleterious mutation in an ovarian cancer predisposition gene based on personal and family history of disease, as the group of interest for our primary analysis. We expect that 65% (n = 2375) of the 3650 participants with personal and/or family history of disease will follow through with providing a saliva sample for analysis, with 75% of those subjects completing the study through the 3 month assessment (n = 1780), which equates to about 450 participants on each of the four study arms.

Each of the three experimental study arms will be compared to the control study arm using a one-sided t-test, with the null hypothesis for each of the three tests that the mean stress score in the experimental arm is more than 4 points above the mean stress score in the control arm. A one-sided significance level of 0.025 is targeted across all three tests, resulting in a Bonferroni-corrected significance level of 0.0083 for each test. The sample size was chosen to achieve 93% power for each test, assuming a standard deviation in the scores of 15.3 based on Schwartz et al., yielding an overall power of 80%.

Our secondary outcome is the “completion rate”, defined as a participant progressing through the entire process from pre-test genetic education/counseling to receiving their test results and post-test genetic education/counseling. We will test two null hypotheses regarding the completion rate: that the completion rate for the participants receiving only online genetic education before testing (Arms A and B) is more than 6% less than it is for participants who were randomized to receive telephone genetic counseling before testing (Arms C and D), and that the completion rate for the participants receiving only online genetic education after testing (Arms A and D) is more than 6% less than it is for participants who were randomized to receive telephone genetic counseling after testing (Arms B and D). This 6% non-inferiority margin is more conservative than the 7.5% margin used by Schwartz et al. [5].

In addition to testing our primary and secondary hypotheses, we will summarize for each study arm the key outcomes related to cognition, affect and behavior, and the changes in scores from baseline. We will also use regression models to model the scores (and changes in the scores from baseline) as a function of study arm, risk group, and other covariates.